Astronomy & Astrophysics manuscript no. C2Hastroph
© ESO 2009
August 8, 2009
C 2 H in prestellar cores
M. Padovani 12 , C.M. Walmsley 2 , M. Taialla 3 , D. Galli 2 and H.S.P. Muller 4
1 Universita di Firenze, Dipartimento di Astronomia e Scienza dello Spazio, Largo E. Fermi 2, 1-50125 Firenze, Italy
2 INAF-Osservatorio Astrotisico di Arcetri, Largo E. Fermi 5, 1-50125 Firenze, Italy
e-mail: padovani@arcetri . astro . it , walmsley@arcetri . astro . it , galli@arcetri . astro . it
3 Observatorio Astronomico Nacional, Alfonso XII 3, E-28014 Madrid, Spain
e-mail: m.tafalla@oan.es
4 I. Physikalisches Institut, Universitat zu Koln, Ziilpicher StraBe 77, 50937 Koln, Germany
e-mail: hspm@phl.uni-koeln.de
Received jdate^, / Accepted idate^
ABSTRACT
Aims. We study the abundance of C2H in prestellar cores both because of its role in the chemistry and because it is a potential probe
of the magnetic field. We also consider the non-LTE behaviour of the N=l-0 and N=2—\ transitions of C2H and improve current
estimates of the spectroscopic constants of C2H.
Methods. We used the IRAM 30m radiotelescope to map the JV=l-0 and N=2—l transitions of C2H towards the prestellar cores L1498
and CB246. Towards CB246, we also mapped the 1.3 mm dust emission, the 7=1-0 transition of N 2 H + and the 7=2-1 transition of
C ls O. We used a Monte Carlo radiative transfer program to analyse the C 2 H observations of L1498. We derived the distribution of
C 2 H column densities and compared with the H 2 column densities inferred from dust emission.
Results. We find that while non-LTE intensity ratios of different components of the N=l—0 and N=2— 1 lines are present, they are of
minor importance and do not impede C 2 H column density determinations based upon LTE analysis. Moreover, the comparison of our
Monte-Carlo calculations with observations suggest that the non-LTE deviations can be qualitatively understood. For extinctions less
than 20 visual magnitudes, we derive toward these two cores (assuming LTE) a relative abundance [C 2 H]/[H 2 ] of (1.0 ± 0.3) x 10~ 8
in L1498 and (0.9 ± 0.3) x 10~ 8 in CB246 in reasonable agreement with our Monte-Carlo estimates. For L1498, our observations in
conjunction with the Monte Carlo code imply a C 2 H depletion hole of radius 9 x 10 16 cm similar to that found for other C-containing
species. We brieHy discuss the significance of the observed C 2 H abundance distribution. Finally, we used our observations to provide
improved estimates for the rest frequencies of all six components of the C 2 H (1-0) line and seven components of C 2 H (2-1). Based
on these results, we compute improved spectroscopic constants for C 2 H. We also give a brief discussion of the prospects for measuring
magnetic tield strengths using C 2 H.
Key words. ISM: abundances - ISM: clouds - ISM: molecules - ISM: individual objects (L1498, CB246) - radio lines: ISM -
Physical data and processes: molecular data
1. Introduction
The ethynyl radical C2H is a crucial intermediate in the inter-
stellar chemistry leading to long chain carbon compounds. It has
a 2 £ ground state giv ing rise to non-negligible Zeeman splitting
dBel & Leroy II1998I) . These facts make it of interest to explore
its abundance variations in nearby prestellar cores. In particular,
it seems worthwhile to test for its presence or ab sence in nearby
cores where CO appears to be depleted (see iBergin & Tafallal
I2007I) . In most circumstances, CO is the main reservoir of C
and thus depletion of CO seems likely to cause a reduction in
the formation rate of C -containing species. However, recently
iHily-Blant et"aT1 d2008h have shown that in L1544 and L183, CN
appears to be present in the high density regions where CO is de-
pleted. This suggests that in some cases, CO depletion may be
"modest" in the sense that it causes the high density core nu-
cleus to be invisible in CO isotopomers but the depletion is not
surBcient to substantially change the abundances of minor C-
containing species. In view of these facts, it seems reasonable to
examine the behaviour of other C-containing species in cores.
C2H is also of interest because hyperhne and spin interac-
tions cause the rotational transitions to split into as many as 1 1
components which can be observed simultaneously with modern
autocorrelation spectrometers. This allows for a precise exami-
nation of the deviations from LTE in individual rotational levels
and eventually an evaluation of the relative importance of colli-
sional and radiative processes. One can observe for example in
the N= 1 -0, 3 mm, transition six components with line strengths
varying by over an order of magnitude and whose relative in-
tensities can be compared extremely precisely. Departures from
LTE have long plagued column density estimates in molecular
lines which are often uncertain by a factor of order 2 as a con-
sequence. Reducing such uncertainties can only be achieved by
understanding the causes of non-LTE behaviour in level popu-
lations and this requires both theoretical (calculations of colli-
sional rates) and observational work. In this study, we attempt to
delineate the problem in the case of C2H from an observational
point of view.
We have chosen to study two cores with contrasting prop-
erties. L1498 is a well studied core in the Taurus complex at a
distance of 140 parsec with a clear CO de pletion hole. I t s den-
sity structure has been studied in detail by Shir ley et alj (2005)
who conclude that their results are consistent with a "Bonnor-
Ebert sphere" of central density 1 - 3 x 10 4 cirr 3 and by Tafalla
et al. (2004, 2006) who find a consid erably higher central den-
sity of 9x 10 4 cm~ 3 . lKirk etaTJ d2006l) used SCUBA polarisation
2
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
measurements to infer a surprisingly low yalue of the magnetic
neld of 10 + 7 fiG in the plane of the sky. lAikawa et al.l 02005)
modelled the molecular distribution and concluded that their re-
sults were consistent with the contraction of a "near equilibrium"
core. There is in general evidence for depletion of C-bearing
speci es such as c-C 3 H 2 and C2S in a central hole of radius 10 17
cm dTafallaet al.ll2006l) . This is in contrast to NH 3 and N 2 H +
which show no signs of depletion in the central high density re-
gion of the core.
CB246 (L1253) is a relatively isolated globule without an as-
sociated IRAS source at a distance of 140-300 parsec (Dame et
al. 1987, Launhardt & Henning 1997) in the general direction of
the Cepheus flare. For the purpose of this study, we adopt a dis-
tance of 200 pc. CB246 is apparently a double core seen both in
NH 3 and C 2 S (Lemme et al. 1996, Codella & Scappini 1998) on
a size scale of roughly 0.1 parsec. T he mass (dependent on the
distance) is in the range 0.2-1 M Q dCodella & Scappinilll998l)
from the NH 3 maps, though there is considerable uncertainty in
this. From a chemical point of view, it is interesting that there is
rough general agreement between the spatial distributions seen
in NH 3 and C 2 S. One aim of the present observations has been
to check if C 2 H shows signs of depletion towards the peak emis-
sion seen in NH 3 .
In this article, we present IRAM 30m maps of the emission
in the N=l-0 transition of C 2 H towards CB246 and L1498. We
supplement this with measurements at selected positions of the
N-2-l transition of C 2 H as well as maps of the 1.3 mm dust
emission, the 7=1-0 transition of N 2 H + and the 7=2-1 tran-
sition of C 18 towards CB246. In Section 2, we describe our
observational and data reduction procedures and in Section 3,
we summarise the observational results from the line measure-
ments. In Section 4, we attempt to use our astronomical observa-
tions to estimate rest frequencies for the individual components
of C 2 H(l-0) and (2-1) and an updated set of hyperAne param-
eters. In Section 5, we give our conclusions concerning the de-
viations from LTE populations for C 2 H as well as a very tenta-
tive interpretation. In Section 6, we give our column density and
abundance estimates in the two objects and discuss the evidence
for depletion of C 2 H . In Section 7, we discuss our results and in
Section 8, we summarise our conclusions.
2. Observations
2.1. C 2 H in L1498 and CB246
The observations were carried out with the IRAM 30m tele-
scope. The C 2 H(l-0) multiplet (at 87 GHz: see TableHJl was
observed at different epochs between August 2007 (only com-
ponent 2, 3, 4 and 5) and July 2008 (all the components), with
1-2 mm precipitable water vapor (pwv). Observations of the
C 2 H(2-1) multiplet (at 174 GHz: seeTable|5]l were performed in
July 2008 with 3-4 mm pwv. The HPBW at the C 2 H(l-0) and
C 2 H(2-1) frequencies are 28" and 14", respectively. In August
2007, the VESPA autocorrelator was used to obtain 10 kHz
channel spacing (corresponding to about 0.034 km s~') with 20
and 40 MHz bandwidths, while in July 2008, the autocorrelator
was used to obtain 20 kHz channel spacing with 20 and 40 MHz
bandwidths for C 2 H(l-0) and 40 kHz channel spacing with 40
and 80 MHz bandwidths for C 2 H(2-1). In this way we covered
the six hyperAne structure (HFS) components of C 2 H(l-0) and
seven HFS components of C 2 H(2-1).
The two cores, L1498 and CB246, were mapped in
C 2 H(l-0) in 2008 in raster mode with a spacing of 20" (chan-
nel spacing 20 kHz) for L1498 and of 15" for CB246 (thus close
to Nyquist sampling for CB246). In 2008, we also observed the
C 2 H(2-1) line at the positions given in Table[2] Finally in 2007,
we observed the C 2 H(l-0) line towards the (0,0) offset with 10
kHz resolution in both sources.
The observed strategy was identical for all measurements:
frequency-switching mode with a 7.5 MHz throw and a phase
time of 0.5 s, with a calibration every 10 to 15 minutes. The data
were reduced using CLASS, the line data analysis program of
the GILDAS softwarfl Instrumental bandpass and atmospheric
contributions were subtracted with polynomial baselines, before
and after the folding of the two-phase spectra. The final rms,
in the main-beam temperature (7 m b), in each channel of width
8v = 0.067 km s" 1 is cr T ~ 50 mK for both L1498 and CB246
and for both the A^=l-0 and the N-2-1 transitions, while the
system temperatures are 7 sys ~ 130 K for the A^=l-0 transition
and 7 sys ~ 1000 K for the N-2-1 transition. In what follows,
all temperatures are on the main-beam scale, 7 m b = F e sT*J B e s,
where T* k is the antenna temperature corrected for atmospheric
absorption, and the forward and beam efficiencies are respec-
tively F e g - 0.95 and B eff = 0.77 concerning the N=l-0 transi-
tion and F eff - 0.93 and B eff = 0.65 for the N-2-l transition.
2.2. N 2 H + (1-0) and C li O(2-1) in CB246
Observations of the N 2 H + (l-0) multiplet (at 93 GHz) and
C ls O(2-l) (at 219 GHz) in CB246 were carried out simulta-
neously in July 2008, with 3-4 mm pwv. The HPBW at the
N 2 H + (l-0) and C ls O(2-l) frequencies are 26" and 11" respec-
tively. The VESPA autocorrelator was used to obtain 10 kHz
channel spacing with 40 MHz bandwidth for N 2 H + (l-0) and 20
kHz channel spacing with 40 MHz bandwidth for C ls O(2-l).
We observed using the frequency-switching mode with a 7.5
MHz throw forN 2 H + (l-0) and a 15 MHz throw for C ls O(2-l)
and a phase time of 0.5 s with a calibration every 10 to 15 min-
utes. We observed a region of 2.5 arc minutes squared in ex-
tent with a spacing of 15". Thus, the N 2 H + map is essentially
Nyquist sampled whereas C 18 is under-sampled (though the
maps shown subsequently are smoothed to a resolution of 26").
For N 2 H + , the final rms, in main-beam temperature units
(7 m b), in channels of width 8v - 0.031 km s _! was cr T ~ 110
mK, with a 7 sys ~ 160 K. For C I8 0, the final rms (channels
of width 8v - 0.027 km s _1 ) was cr T ~ 200 mK, with a
7 sys ~ 400 K. The forward and the beam efficiencies are re-
spectively F eff = 0.95 and fi eff = 0.77 for N 2 H + and F eff = 0.91
and B eff = 0.55 for C ls O.
2.3. Bolometer map of CB246
CB246 was observed in the 1.3 mm continuum with the IRAM
30m telescope in May, November and December 2007. We used
the MAMB02 117-channel bolometer array in the on-the-fly
mode with a scanning speed of 6" s _1 , a wobbler period of 0.5
s and a wobbler throw of 70", reaching an rms of 3 mJy/beam.
The bolometer central frequency is 250 GHz and the half power
bandwidth (HPBW) at 1.3 mm is approximately 11".
3. Observational results
3.1. C 2 H(1-0)
In Fig. [T| we show our map of the integrated intensity in the
C 2 H(l-0) (J',F' J,F = 3/2,2 1/2, 1) transition su-
1 http://www.iram.fr/IRAMFR/GILDAS.
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
3
perposed on the map of the 1.3 mm dust continuum emission
smoothed at 28" together with the positions observed in the
N-2-1 transition. One sees that towards L1498, while the dust
emission has a single peak at the centre of the map, the C2H
emission has a broad "plateau" with an extent rather similar to
that of the dust, showing two peaks to the SE and NW of the dust
emission peak. Rather similar distributions have been seen in
several other C-bearing species (e.g. CS, C2S, CO and C-C3H2,
see Tafalla et al. 2002 and Tafalla et al. 2006) and has been at-
tributed to depletion onto grain surfaces in the dense gas asso-
ciated with the dust continuum peak. Our data suggest that C2H
behaves in a similar fashion towards L1498.
The situation is rather different towards CB246 (see Fig. |2j
upper panel) where we see that the dust and C2H emission have
rather similar distributions. Both are double peaked, although
the C2H NW emission peak is offset about 30" to the south of
its counterpart in dust emission whereas towards the SE peak the
difference is marginal.
50 I
o
O
w
o
u
(0
<
-100
100
Aa. [arcsec]
-100
Fig. 1: Dust emission at 1.3 mm (after smoothing to 28",
coloured map) towards L1498, from Tafalla et al. (2002), super-
posed to the emission of C 2 H(l-0) (J',F' -> J,F = 3/2,2 -»
1 /2, 1), black contours . Contours represent40, 50, 60, 70, 80, 90
and 95 per cent of the peak value which is 0.67 K km s . The
scale bar to the right of each panel gives the continuum fiux in
mJy/beam. The crosses indicate the positions used for the cuts
in Fig.[3]while the circles show positions observed in C2H(2-1).
The (0,0) position corresponds to »(2000) = 04 h 10 m 51.5 s ,
(5(2000) = 25°09'58".
It is interesting to compare the distributions of different com-
ponents. We do this in Fig. [3] where we compare cuts in all
the 1-0 component lines with cuts in the continuum intensity.
One sees that ratios of different pairs of components do not vary
greatly along these cuts, although the ratio of the strong 87316
MHz component (no. 2 in Table@]l to the weak 87284 MHz line
(no. 1) at the L1498 dust peak is somewhat smaller (about 45%)
than elsewhere in the cut. However, this ratio is always of order 3
as compared to the expected value of 10 for optically thin emis-
sion (suggesting a moderately optically thick 87316 MHz line).
There are thus indications of saturation of the stronger compo-
nents of the C2H(l-0) line and we conclude that optical depth
effects are present but lines have moderate opacities, as shown in
TableQ]for selected positions in the two observed sources. It is
o
0)
w
o
<a
<
-50
o
0)
w
o
u
<a
•o
<
-50
Aa [arcsec]
-100
Fig. 2: Upperpanel: dust emission at 1.3 mm (after smoothing to
28", coloured map) towards CB246 (see Sect. 12.31 1. superposed
to the emission of C 2 H(l-0) (J',F' -> J,F = 3/2,2 -> 1/2, 1);
circles show positions observed in C 2 H(2— 1); asterisk shows the
position of the 2MASS object (see Sect. 13.41 ). Lowerpanel: dust
emission at 1.3 mm (after smoothing to 26", coloured map) to-
wards CB246 superposed to the emission of C ls O(2-l), black
contours, andN 2 H + (l-0) (F[,F' F\,F = 1,2 —> 0, 1), white
contours; Both panel: contours represent 40, 50, 60, 70, 80, 90
and 95 per cent of the peak values which are 0.67, 0.70 and 0.39
K km s 1 for C 2 H(l-0), C lg O(2-l) and N 2 H + (l-0), respec-
tively; the crosses indicate the positions used for the cuts in Fig.
[3]and|6l solid triangles represent the peaks of NH3(1,1), from
Lemme et al. (1996), and solid sauares the peaks of C 2 S(2i-
lo), from Codella & Scappini (1998). The scale bar to the right
of each panel gives the continuum flux in mJy/beam. The (0,0)
position corresponds to «(2000) = 23 h 56 m 43.6 s , (5(2000) =
58°34'09".
noticeable also that while C 2 H has a minimum towards the dust
peak in L1498 (and we conclude this is a real column density
minimum), there is a little indication of variation in the ratio of
C 2 H to dust continuum intensity crossing the SE peak in CB246.
Another indication of line transfer effects can be obtained
from comparing line prohles of the different C^H^l-O) compo-
nents as shown in Fig.|4]where we show the high (10 kHz) spec-
tral resolution data from 2007. It is noticeable that towards the
dust peak, the weakest 87407 MHz component (no. 5 in Table
[4]l has a maximum at a velocity where the strong 87316 MHz
4
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
Table 1: Line parameters fl observed in C2H(l-0)
a, S offsets
Vlsr AV
b
T
total
L . J
[km s~'] [km s -1 ]
L1498
0,0
7.80(0.01) 0.17(0.01)
29.4(0.3)
60, -40
7.74(0.03) 0.21(0.01)
28.3(0.4)
CB246
0,0
-0.83(0.01) 0.24(0.01)
13.5(0.3)
15,-30
-0.78(0.01) 0.26(0.01)
20.9(1.1)
-60, 15
-0.87(0.01) 0.27(0.01)
14.4(0.7)
" Numbers in parentheses represent the errors from the fit.
b Sum of the peak optical depth of the six hyperhne components.
[L1498] NW-SE cut
[CB246] EW cut of the SE peak
20 -20
Aa [arcsec]
Fig. 3: Upper panels: Continuum emission flux measured along
the NW-SE cut in L1498, see Fig.Q] and along the EW cut of the
SEpeak in CB246, see upperpanel of Fig. [2] Lowerpanels: ratio
between the integrated intensity of the different hyperfine com-
ponents of C2H(l-0) with the continuum emission flux in the
same positions of the upperpanels. The typical error on the ratio
is 0. 16 K km s" 1 mJy~' beam (component 1, solid magenta line
with triangles; component 2, solid red line with sguares; com-
ponent 3, solid blue line with diamonds; component 4, dashed
blue line with diamonds; component5, solid green line with pen-
tagons; component 6, dashed magenta line with triangles; com-
ponent labels are given in Table|4|.
line (no. 2) shows a dip between two peaks. This is the signature
expected for "self- absorption" by a foreground layer of den-
sity lower than that responsible for the bulk of the emission. The
effect however is not surHciently strong to change the above con-
clusions concerning the importance of optical depth effects. We
note moreover that towards CB246, no effects of this type are
seen.
3.2. C 2 H(2-1)
The C2H(2-1) line has had little (if any) attention and it was
thus interesting that we succeeded in detecting 7 of the 1 1 com-
ponents of the line in both sources. Line parameters are given in
Table|2]and sample proAles are shown in Fig. [5] It is interest-
ing (see discussion below) that within the errors, line intensities
are consistent with LTE but that, in particular towards L1498,
the total optical depth derived from an LTE fit (see Table [2]) is
large and corresponds to an optical depth of 3.1 in the strongest
Fig. 4: Component 2 (black), 3 (green), 4 (blue) and 5 (red) of
C 2 H(l-0) observed at high resolution in 2007 at the offset (0,0)
in L1498 (left panel) and CB246 (right panel). In order to com-
pare line shapes and intensities, components have been shifted
in frequency. Component labels and rest frequencies are given
in column 1 and 3 of Table[4]i.
174663 MHz component (no. 2 of Table[5]). We note however
that slight errors in rest frequencies could inAuence this inter-
pretation.
7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4
v [km s -1 ]
-1.0 -0.5
v [km s"']
0.0
Fig. 5: Hyperfine components of C2H(2-1) observed at the offset
(0,0) in L1498 (left panel) and CB246 (right panel). In order
to compare line shapes and intensities, components have been
shifted in frequency. Component labels and rest frequencies are
given in column 1 and 3 of Table|5]l.
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
5
[CB246] SN out of the SE peak
[CB246] SN cut of the NW peak
Table 2: Line parameters" observed in C2H(2— 1)
a, S offsets
Vlsr
AV
b
T
total
L , J
LKm s J
[Km s J
L1498
0,0
7.84(0.01)
0.23(0.01)
8.6(1.8)
60, -40
7.76(0.01)
0.25(0.01)
9.7(1.9)
CB246
0,0
-0.84(0.01)
0.35(0.02)
1.5(1.2)
15,-30
-0.73(0.01)
0.30(0.01)
5.4(1.0)
-60, 15
-0.88(0.01)
0.30(0.01)
5.5(1.2)
" Numbers in parentheses represent the errors from the fit.
* Sum of the peak optical depth of the seven hyperhne components.
3.3. C l *OandN 2 H +
In the lower panel of Fig. [2] we show a superposition of our 1.3
mm continuum map (smoothed to an angular resolution of 26")
with the C 18 0(2-l) and the isolated component of N 2 H + (l-0)
(F'F r
F],F = 0,1
1,2) integrated intensity maps, to-
gether with the emission peaks in NH3(1,1), using the data ob-
tained by Lemme et al. (1996) and in C2S(2i - lo) by Codella
& Scappini (1998). There is a general similarity between the
distributions of C 18 seen here and that of C2H (Fig. |2j up-
per panel). In general, also the C 18 distribution resembles that
seen in the continuum but there are clear shifts (for example in
the SE) between the peaks seen in the dust emission and in C 18 0.
In the NW, also the "bar"-like structure seen in the continuum to
be extended N-S is also present in C ls O but seems even more
elongated (dimensions of 100" in the continuum as compared to
120"inC 18 O).
To the SE there is reasonable agreement between the con-
tinuum and N 2 H + intensity distributions and the ammonia peak
appears to be consistent with these. To the NW on the other hand,
the continuum is extended in a "bar"-like feature and N 2 H + (but
also C2H) peaks at the south end of this 40" to the south of the
continuum peak. These differences suggest chemical differenti-
ation in CB246 albeit on a scale smaller than in L1498. Besides,
the N2H + peaks resemble the continuum structure in the SE but
differ considerably in the NW suggesting that in this case (in
contrast to L1498 and L1544, see Tafalla et al. 2006), the N 2 H +
abundance may be varying with position. The shift between the
N2H + NW peak and the continuum peak is unusual, but not un-
precedented (see Pagani et al. 2007). However, a more detailed
study of the radiative transfer and excitation is needed to conhrm
this.
In Fig. [6] we compare cuts in the strongest component of
C 2 H(l-0) (component 2, see Table|4|, C 18 0(2-l) and the iso-
lated component of N 2 H + ( 1 -0) with cuts in the continuum inten-
sity. N 2 H + and C 18 show a rather constant line-to-continuum
ratio in the SE peak cut, while in the NW peak cut, these
molecules do not look like the continuum structure, especially
the C 18 emission which appears to be shifted down compared
to the continuum emission.
3.4. Mass of CB246
In the continuum, CB246 shows a double-peaked prohle, with a
rounded SE clump with a radius (measured at half-power con-
tour) of 33" (corresponding to 0.03 pc at the distance of 200 pc)
and a more elliptical NW clump with a major and a minor axes of
100" (orO.l pc) and 44"(or 0.04 pc), respectively. Dust emission
is generally optically thin at millimetre wavelengths and hence
is a direct tracer of the mass content of molecular cloud cores.
50
Ao [arcsec]
Fig. 6: Upper panels: Continuum emission flux measured along
the SN cut of the SE and the NW peaks in CB246, see upper
panel of Fig.|2] Lowerpanels: ratio between the integrated inten-
sity of the component 2 of C 2 H(l-0), red sguares, C ls O(2-l),
blue diamonds, and the isolated component of N 2 H + (l-0), ma-
genta triangles, with the continuum emission flux in the same
positions of the upperpanels. The typical errors on the ratio are
0.16, 0.15 and 0.08 K km s" 1 mJy 1 beam for C 2 H, C ls O and
N 2 H + respectively.
In the hypothesis of an isothermal dust source, the total (dust
plus gas) mass is related to the millimetre flux density, S 1.3, in-
tegrated over the solid angle £2beam according to the following
equation
M =
S 1.3 d 2
Ki.3Bi.i(Td)
(1)
where Bu,(Td) is the Planck function calculated at the dust tem-
perature T& = 10 K and K1.3 is the dust opacity per unit mass
which is assumed constant and equal to 0.005 cm 2 g _I for
prestellar cores of intermediate densities (n < 10 5 cm -3 ), as-
suming a gas-to-dust mass ratio of 100 (Henning et al. 1995,
Preibisch et al. 1993). We found a FWHM mass of 1.2 M for
both clumps and a total mass of 3.3 M Q which has been evaluated
taking into account the continuum flux above two times the rms.
The average column and number density of molecular hydrogen
have been evaluated within the same area used to compute the
mass and are to N(H 2 ) = 1 .9 x 10 22 ctrT 2 and n(H 2 ) = 7.6 x 10 4
cm~ 3 , respectively.
To the south of the NW emission peak, there is a bright in-
frared source (2MASS J23563433+5834043, see Fig.|2) which
appears to be a heavily reddened background star with visual ex-
tinction of ~ 20 magnitudes, consistent with our column density
estimates.
3.5. Gas kinematics
From the hyperAne fit of C 2 H(l-0), using the rest frequencies
discussed below (see Sect. 14. lj , we derived the line center ve-
locities, Vi sr , and in Fig. [7] we present the radial proAles which
are almost flat on the average, especially for L1498. Radii are
computed with respect to the ofFset (0,0) for the two sources (see
Fig.[TJand|2).
Following Goodman et al. (1993), we checked for the pres-
ence of a velocity gradient, across these two cores using
6
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
l
1 1 1
i u
■ ■ ■
.-:
1 1 1 1 1 1
■■■■■_ ■
■ ■ «
■ ■>■-■■■
: L1498
— 1 — 1 — \
■ 1 1 —
;!
■
\
■ ■ ■■
■ ■ ■
1
■
:*■
1 1 1 1 1 1
m
i[ ■■*: ,
:■ "*■ ■■ ■ :
; CB246
1 1
1 ! 1
1
1 1 1 1 1 1
100
50
R [arcsec]
100
Fig.7: Radial profile of line center velocity for L1498 (upper
panel) and CB246 (lower panel) derived from the hyperhne-
structure fits of C2H(l-0). The uncertainty of the velocity es-
timates is of the order of 0.003 km s _I and the dashed line rep-
resents the mean value.
-100
100 o
Aot [arcsec]
-50
Aa [arcsec]
-100
Fig. 8: C^H^l-O) velocity gradient toward L1498 (upper panel)
and CB246 (lower panel): the integrated intensity map of com-
ponent 2 of C2H is shown in gray scale (black contour represents
the 50% of the peak value) while arrows represent the local ve-
locity gradients in the adjacent nine points pointing in the direc-
tion of increasing velocity and with the length proportional to
the magnitude. The arrow in the bottom- right corner represents
the total velocity gradient.
Table 3: Results of gradient htting.
C2H(l-0) data. A further check has been done on CB246 us-
ing N2H + (l-0) and C I8 0(2-1) data, hnding a good agreement
within the errors with the corresponding values from C2H. In
Fig.[8]we show the local velocity gradients in the two cores and
in Table[3] we show our results as regards the velocity gradient,
5f, its direction, and the mean LSR velocity, (Vi S r)- All these
quantities are in good agreement with the estimates of Goodman
et al. (1993) and Tafalla et al. (2004) who observed L1498 using
NH3 and N2H + as a tracer. Besides, we evaluated the angular
velocity, 10, assuming that the angular velocity vector points in
the direction given by a>, that is a> = (ĕf/ sin i) a>, where i is the
inclination of a> to the line of sight and the position angle of a>
is given by ©d, = 0g? + n/2. Statistically, for a random distri-
bution of orientation, (sin i) = 4/n (Chandrasekhar & Miinch
1950, Tassoul 1978). To quantify the dynamic role of the rota-
tion in a cloud, we calculated the ratio between the rotational and
the gravitational energy, denoted with /3. For a uniform density
sphere,j8 = (a> 2 R y )/(3>GM), where R is the rotation radius, G the
gravitational constant and M the mass, and the specihc angular
momentum is dehned as L/M = (2/5)R 2 u>.
The values of /3 we found suggest that the clouds have lit-
tle rotational energy as generally found in molecular cloud cores
(e.g. Goodman et al. 1993). The specihc angular momentum we
evaluated for these two cores is of order 10 -3 km s _1 pc. In the
hypothesis that a core has been formed from a parental clump
with equal mass and at the same galactocentric distance of the
core (about 7 kpc for L1498 and CB246), we computed a ro-
tation frequency of order 10~ 15 s _1 assuming that the angular
momentum of the clump is due to Galactic differential rotation
(Clemens, 1985). This value corresponds to a specihc angular
momentum of 3 x 10~ 2 km s" 1 pc, that is 30 times greater than
the L/M of the cores. This shows that the specihc angular mo-
mentum of the cores is low compared to larger (galactic) scales.
Thus angular momentum is lost in forming cores as expected by
Ohashi (1999) for cores with radius greater than 0.03 pc. In par-
ticular, the cores lies on the relation founded by Ohashi (1999)
for L/M as a function of the radius.
quantity
unit
L1498
CB246
<$
[km s~' pc -1 ]
0.56 ± 0.05
0.65 ± 0.09
@cg
[degEofN]
-37 ± 1
143 ±4
(V hr )
[km s-']
7.79 ± 0.04
-0.83 ± 0.05
LO
[10- 14 s- 1 ]
2.3 ±0.2
2.7 ± 0.4
P
[10- 3 ]
-2.5
~ 4.0
L/M
[10- 3 km s- 1 pc]
1.0 ±0.2
1.2 ±0.4
4. Evaluation of new spectroscopic parameters for
C 2 H
The C2H radical is a linear molecule. Its unpaired electron
causes a splitting of each rotational level into two fine structure
levels. In addition, the spin of the H nucleus causes each fine
structure level to be split further into two hyperhne levels. The
strong transitions are those having AA^ = AJ = AF which leads
to four strong transitions at higher quantum numbers. However,
transitions with AN + AJ or with AJ + AF, which have essen-
tially vanishing intensities at high jV, may have fairly large rel-
ative intensities at lower values of A^; the only restriction which
has to be followed strictly is AF = or ±1. In the case of the
^=1-0 and N=2-l transitions, this leads to 6 and 1 1 hyperhne
lines, respectively, with non-zero intensity.
4. 1 . Rest frequencies
Preyious rotational data of C2H have been summarised by
Mulleretal. (2000). They had measured submillimetre transi-
tions up to 1 THz; the small hydrogen hyperhne splitting was
not resolved. Their data set also included 6 hyperhne lines
of the N=l-0 ' meas ured toward a position north of Orion-KL
dGottlieb et al.lll98"3l) . four hyperhne components of the N=2- 1
transition obtained in the laboratory by the same autho rs as well
as four N=3-2 hyperhne features dSastry et alj Il98l1) . These
data as well as predictions are available from the recommended
CDMS0 (Cologne Database for Molecular Spectroscopy, Miiller
etal. 2001,2005).
2 http : //www . astro . uni -koeln . de/cdms/
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
7
In the present investigation we have observed all six HFS
components of the N=l—0 transition as well as seven compo-
nents of the N-2-l transition. Because of good to very good
signal-to-noise (S/N) ratios of all these lines and because of
the small line width it seemed promising to determine im-
proved rest frequencies for these lines with respect to those
available in the CDMS. The main sources affecting the accu-
racy of rest frequencies, besides S/N and line width are expected
to be the symmetry of the (usually Gaussian) line shapes and
the accuracy with which the velocity structure of the source is
known. On th e basis of the wel l determined res t frequencies
ofN 2 H + (l-0) dPagani et alj|2009h and C ls O(2-l) dMiiller et alj
l2001h we have derived at the dust continuum peak an LSR ve-
locity of 7.80 km s _1 f or L1498 based on the observations of
these two transitions by Tafa lla etall (|2002) and an LSR veloc-
ity of -0.83 km s _1 for CB246 from the present observations
of these transitions (see Sect. 12.21 ). Employing these LSR veloc-
ities, rest frequencies of C 2 H have been determined separately
for the two observed sources. They differ on average by 3 and
17 kHz, respectively, for the N=\ — and N=2-l transition and
the averages for each HFS component are given in Table|4]and[5]
respectively together with the assignments, the estimated uncer-
tainties, the residuals o-c between the observed frequencies and
those calculated from the final set of spectroscopic parameters,
and the relative intensities /.
Table 4: Observed hyperhne structure components of the /V' — >
N = 1 — > transition in C 2 H: rest frequencies", residuals o-c
between observed and calculated frequencies 6 , and relative in-
tensities /.
comp.
Transition
V
o-c
/
no.
J',F' -
J,F
[MHz]
[kHz]
1
3/2,1 ->
1/2, 1
87284.105 (10)
-3
0.042
2
3/2,2 ->
1/2, 1
87316.898(10)
4
0.416
3
3/2, 1 ->
1/2,0
87328.585 (10)
-2
0.207
4
1/2, 1 ->
1/2,1
87401.989(10)
-3
0.208
5
1/2,0 ->
1/2,1
87407.165(10)
8
0.084
6
1/2, 1 ->
1/2,0
87446.470(10)
0.043
8 Numbers in parentheses are one standard deviation in units of the
least signincant ngures.
* Calculated from present set of spectroscopic parameters in Table|6]
Table 5: Observed hyperhne structure components of the N' — >
N = 2 — > 1 transition in C 2 H: rest frequencies", residuals o-c
between observed and calculated frequencies'', and relative in-
tensities /.
comp.
Transition
V
o-c
/
no.
J',F' -> J,F
[MHz]
[kHz]
1
5/2,2 -> 3/2,2
174634.861 (16)
6
0.017
2
5/2,3 -» 3/2,2
174663.199(12)
0.350
3
5/2,2 -» 3/2, 1
174667.629(14)
-12
0.232
4
3/2,2 -> 1/2, 1
174721.744(14)
-8
0.197
5
3/2,1 -> 1/2,0
174728.071 (12)
12
0.083
6
3/2, 1 — » 1/2, 1
174733.210(14)
-15
0.052
7
3/2,2 -> 3/2,2
174806.843 (14)
-7
0.045
" Numbers in parentheses are one standard deviation in units of the
least signiticant ngures.
* Calculated from present set of spectroscopic parameters in Table|6]
4.2. Spectroscopic parameters
These rest frequencies, together with p reyious laboratory values
dSastryet al.l 1 198 lt iMuller et al1l2000l) were used to calc ulate
spectroscopic parameters for C 2 H. As in lMiiller et alJd2000l) . the
rotational constant B, the quartic and sextic distortion terms D
and H, the electron spin-rotation parameter j along with its dis-
tortion correction j D , as well as the scalar and tensorial electron
spin-nuclear spin coupling terms bf and c were determined. The
distortion correction b°. to bf was not determined with signih-
cance, e ven though it s absol ute uncertainty was slightly smaller
than in Mtill er etail d2000). and it did not contribute signih-
cantly to the reduction of the rms error. It was thus omitted
from the final fit. However, it was found that the 'H nuclear
spin-rotation constant C impnwed the quality of the fit by a non-
negligible amount of 12 % even though it was barely determined.
As, in addition, its value appeared to be reasonable, see further
below, this constant was retained in the final fit. The resulting
spectroscopic parameters are g iyen in T able|6]together with the
most recent previous values by Mii ller et alj (120001) .
Table 6: Spectroscopic parameters" (MHz) of C 2 H in compari-
son to previous values.
Parameter
present value
Miiller et al. (2000)
B
43674.5177(13)
43674.5289(12)
D x 10 3
105.515(53)
105.687 (51)
Hx 10 6
-0.42 (32)
0.32(32)
7
-62.6103(45)
-62.6029 (43)
y D X 10 3
-2.227 (250)
-2.313(255)
b F
44.4788 (82)
44.4922(183)
-0.0100(38)
C
12.2389 (232)
12.2256(261)
c
-0.0087 (53)
Numbers in parentheses are one standard deviation in units of the
least signincant tigures.
The most striking feature in the co mparison of the pres ent
C 2 H rest frequencies with those from iGottlieb et alj d 19831) is
that the latter are on average 28.6 kHz higher in frequency, rang-
ing from complete agreement to 5 1 kHz for the individual HFS
components. Moreove r, the deyiations are a few times the un-
certainties reported by Gottl ieb et alj d 19831) for some lines. In
fact, employin g the presen t spect roscopic parameters given in
Table|6] the Gottli eb et alJ d 19831) data are reproduced to only
3.6 times their reported uncertainties on average. We suspect
that the LSR yelocit y of the source north of Orion-KL used in
iGottlieb et alj dl983b was not as well known as the authors as-
sumed and that for the laboratory measurements the frequency
determinations were slightly off or the estimates of the uncer-
tainties were slightly too optimistic.
The present rest frequencies have been reproduced to 0.6
times the uncertainties, slightly better still for the N=l-Q tran-
sition, suggesting that the uncertainties in Table [5] and even
more so in Table |4] have been judged somewhat too conser-
vatively. On the other hand, uncertainties in the LSR velocity
may j ustify such a conser vative error estimate. The N=3-2 lines
from Sast ry et alj dl98jj) have been reproduced to better than
18 kHz, which is much better t han the ~ 45 kHz w ith which
these data where reproduced in Miill er et~ai] d2000i) . In other
words, these N=3— 2 rest frequencies are much better compat-
ible with the pre sent N=l-Q a n d N= 2-l rest frequencies than
with those from Gottlieb et a il d!983h . The submillimetre data
from iMiiller et alT 72000) are at somewhat higher frequencies
8
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
and quantum numbers such that both present and previous spec-
troscopic parameters in Table[6]reproduce these data on average
to about 0.6 times the reported uncertainties.
While frequency deviations of a few tens of kHz may possi-
bly be neglected in observations of hot cores or similar sources,
the deviations are rather considerable for investigations into the
dynamics of dark clouds as in the present study. The deviations
are also rerlected in the small (~ 1 1 kHz) differences in the rota-
tional constant B, see Table[6] Nevertheless, these deviations are
more than four times the combined uncertainties an d thus clearly
signih cant. It is worth noting that the B value in Miiller et al.
(2000) is determined essentially by the data from lGottlieb et aU
dl983b . The slightly higher D value from the present study, just
outside the combined uncertainties, compensates the change in B
to some degree for the submillimetre lines. The sextic term H is
essentially the same as in the previous study, still not determined
with signihcance, but probably of the right order of magnitude.
Changes in the fine structure parameters and in the larger
hyperhne structure parameters are essentially insignihcant.
had been used in the previous
(2000) since its inclusion improved the
The distortion correction b°.
fit of Miiller et al
iGottlieb et al.l (1 19831) data to be reproduced on average from
1.31 times the uncertainties to 0.93 times the uncertainties. In
present trial fits its value was determined as -0.0021 (33) MHz.
The magnitude of the ratio b D /bf is now much closer to that
of y D /y, but still much bigger than that of D/B. However, the
uncertainty was larger in magnitude than the value, and the in-
clusion of the parameter in the fit improved the quality of the
fit only insignihcantly. The term b D was consequently omit-
ted from the final fit. The inclusion of the 'H nuclear spin-
rotation term C in the present fit requires some explanation.
This term is usually small and negative and frequently scal-
ing with the rotational constant is the dominant contribution
to its size. The C2H value of -8.7 (53) kHz agrees within its
large unc ertainty with the experiment al one of -4.35(5) kHz
for HCN (Ebenstein & Muenteri ll984T) and with the calculated
yalues of -4.80 kHz and -5.55 kHz for HCN and HCO +
dSchm id-Burg kĕt~aT1 120041) and appears thus to be reasonable
as these species have fairly similar rotational constants.
5. Non-LTE hypertine populations
In this Section we consider the evidence for non-LTE popula-
tions in the hyperhne levels sampled by our observations. We
first compare the observed intensity ratios of C2H in L1498 and
CB246 with the predictions of simple LTE models and find that
while real deviations from LTE populations are present, the LTE
assumption is an approximation which is useful for many pur-
poses. We then consider a Monte Carlo radiative transfer pro-
gram for the case of C2H and show that the observed deviations
can be approximately understood with an educated guess at the
(unknown) C2H collisional rates.
5.1. Single- and two-layer models
We note first that for the case of LTE between different hyper-
fine levels (i.e. all transitions of a given multiplet have the same
excitation temperature), we expect that for a homogeneous slab,
the ratio of line intensities of two transitions Rij is given by:
where t is the total transition optical depth and f is the rela-
tive line strength of the j-th component as e.g. in Table|4] Thus,
as the optical depth varies, the expected Rjj varies from f/fj to
unity and for two different pairs of transitions, one can derive a
curve in the plane (Ry, /?«). In Fig.[9] we give for various combi-
nations of transitions in the N=\-0 transition our observed line
ratios compared with the result expected on the basis of equation
[2] and in Fig.[10] we show analogous results for the N-2-l line.
We see that the observed intensity ratios in C2H(l-0) are not
consistent with equation[2]although the deviations are not large.
In particular, we note in Fig. [9] that ratios such as R32 (upper
right panel) at some positions attain values below the optically
thin expected value of 0.5. More signihcantly perhaps the ob-
served ratio R34 (bottom right panel) of the 83728 and 87401
MHz lines which have essentially the same line strength is typi-
cally of order 0.8-0.9 and at almost all positions less than unity.
These differences are not signihcant at individual positions, but
appear to be signihcant statistically. The results in Fig.|9]are for
L1498, but similar trends hold for CB246.
Fig. [TOl shows analogous results for C2H(2-1). There are
fewer positions but again, we see signihcant deviations from the
homogeneous single layer prediction of equation[2] The ratio /^32
(upper left panel) of the 174667 and 174663 MHz lines is typi-
cally between 0.5 and 0.6 as compared to the optically thin LTE
prediction of 0.66. These differences are small but appear to be
signihcant.
Equation [2] assumes a homogeneous one-dimensional solu-
tion of the transfer equation and it is clear that for a real prestel-
lar core, gradients in both temperature and density are present.
Indeed observations of some cores in strong ground state tran-
sitions such as HCO + (1-0) are complicated by absorption in a
foreground layer whose excitation appears to be essentially that
of the cosmic background (Tafalla et al. 1998). In the particu-
lar case of C2H towards L1498, the "self-absorption" observed
in the 87316 MHz line (see Fig.|4| suggests that something of
this sort may occur also for C2H . In view of this, we have also
considered a two-layer model of the type discussed by Myers et
al. (1996) with background (b) and foreground (/) layers having
excitation temperatures T e% j, and T ex j and optical depths t/ and
Tb respectively. We then find for the emergent intensity of i-th
component:
T mb ,j = J(T exJ )[\ - e- T f '] + J(T^ b )[\ - e- T '"]e-^
-J(T bb )[\ - e- T »-- T f-]
where
J(T)
exp(T /T) - 1
(3)
(4)
is the Planck-corrected brightness temperature and T bb = 2.73 K
is the temperature of the cosmic background. T = hv/k, where
v is the transition frequency and h and k represent Planck's and
Boltzmann's constants respectively. The optical depth of the k-
layer (k = /, b) for the i-the component is dehned as t^,- = T f<pj,
where f is again the relative intensity and <pj is the prorile which
we assume as gaussian, that is
exp
4[V-(-l)"Wt,;>] 2 ln2
(T
(5)
Rjj =
1 - exp(-fT)
1 - exp(-/)T)
with m = for the foreground layer and m = 1 for the back-
^ ' ground layer. We have included in Fig. [9] and [10] predictions
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
9
based on our two-layer model varying T ex j and we derive the
dashed curves using equation [3] We see from this that the ef-
fect of "foreground layers" is to increase ratios such as /^32 in
some cases to values of order unity. Essentially, this is due to ab-
sorption of the strong 87316 MHz line (component 2). One sees
from both Fig.|9]and[T0]that this is not what is required to fit the
observed data and in fact the two-layer models are a poorer ap-
proximation to the observed intensity ratios than the single-layer
LTE model. We thus conclude that understanding the observed
line ratios requires a proper non-LTE treatment.
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
R 54 R 32
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
R 63 R 34
Fig. 9: Ratio of the integrated intensities of some couples of
components of C2H(l-0) in L1498, where R,j represents the ra-
tio between J T m bjdV and J T m bjdV. Observational data, black
empty sguares; one-layer model, blue solid curve; two-layer
model with difFerent T ex j, green dashed curves; results from
Monte Carlo model, red filled triangles. The typical errors on
the ratios are of the order of 8.5 x 10~ 2 .
5.2. Monte Carlo treatment of radiative transter in the C^H
rotational transitions
The two-layer model, with its assumption of the same excitation
temperature for all hyperhne components in each layer and its
discontinuous jump in conditions between the two emitting re-
gions, misses an important part of the complex pattern of level
populations responsible for the emission in C2H(l-0). As can
be seen in Fig. |H the hyperhne components of C2H(l-0) are
simultaneously subthermal and optically thick in both L1498
and CB246, and under these conditions, the excitation temper-
ature of each component is determined by a delicate balance
between collisions and trapping. This balance will be different
for each transition depending on its particular optical depth, and
it will therefore give rise to differences in excitation between
the components of the N=l—0 multiplet. The large optical depth
of the lines, in addition, makes the emission from each compo-
nent originate in gas at different depth in the core, and this also
contributes to differences in the observed line intensities. Such
complex interplay between excitation and optical depth effects
cannot be treated accurately with the two-layer model, and its
analysis requires a more sophisticated numerical scheme. In this
Section, we present the result of a (partial) solution to the C2H
radiative transfer problem using the same Monte Carlo code used
Fig. 10: Ratio of the integrated intensities of some couples of
components of C2H(2-1), where Rj represents the ratio be-
tweenj Tmb^dyand jT m bjdV. Observational data, blackempty
sauares for L1498 and black filled squares for CB246; one-
layer model, blue solid curve; two-layer model with different
T e x,f, green dashed curves; results from Monte Carlo model for
L1498, red hlled triangles.
in Tafalla et al. (2004, 2006) to analyse the emission from a num-
ber of molecular species in the L1498 and L1517B cores.
As the modelling of the L1498 emission in Tafalla et
al. (2004, 2006) already fixed the physical description of this
core (radial prohles of density, temperature, and velocity), the
only parameter left free to model the C2H emission is the radial
prohle of abundance. Unfortunately, the C2H analysis is limited
due to the lack of known collision rates for the species, and this
forces us to make an educated guess of this important set of co-
efficients. Following Turner et al. (1999), we approximate the
collision rates of C2H using those for HCN calculated by Green
& Thaddeus (1974). From these rates, which do not include hy-
perhne structure, we derive a new set of rates with hyperhne
structure assuming that the new rates are simply proportional
to the degeneracy of the final state (Guilloteau & Baudry 1981,
Lique et al. 2009). Additional AN = rates were included by
following the recipe from Turner et al. (1999), and a total of 30
levels with 37 transitions (up io N -1 and an energy equivalent
to 120 K) were used in the calculation.
Following the analysis of other species in L1498, the goal of
our C2H modeling was to fit simultaneously the combined radial
prohle of A^ =1-0 integrated intensity together with the central
spectra of the different components of the A^=l-0 and N-2-l
multiplets, assuming a uniform kinetic temperature of 10 K (as
derived from a non-LTE analysis of the NH3 data by Tafalla et
al. 2004). A first set of model runs using the collision rates de-
scribed before predicted excitation temperatures that were too
high in the outer core layers, a situation that is inconsistent with
the self absorbed profile seen in the thickest N =1-0 compo-
nent (Fig. |4j). This inconsistency indicated that the guessed C2H
collision rates were too large, and that they should be signif-
icantly reduced in order to match the observations. To avoid
introducing artifacts in the relative excitation of the hyperfine
components, the collision rates were reduced dividing them by
a global factor of 3. With these corrected collision rates, a fit
was achieved by assuming a constant C2H abundance with re-
spect to H2 of 8 x 10~ 9 and a central depletion region of radius
10
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
50 100
R (arcsec)
150
2
1.5
1 l
0.5
2
1.5
1
comp. 1 comp. 2 :; comp. 3
comp. 4
comp. 5
7 7.5 8 8.5 7 7.5
v lsr (km/s)
i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r
comp. 6
3.5 7 7.5 8 8.5
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 [- comp. 1
0.5 -
comp. 4
1 1 1 1 1 1 1 1 1 1 1 1 1 1 ii 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
comp. 2 — comp. 3
comp. 5
comp. 7 -7 7.5 8 8.5 7 7.5 8 8.5
comp.
co
I
7 7.5 8 8.5
v lsr (km/s)
Fig. 11: Comparison between observations and our best fit
Monte Carlo model of the C2H emission in L1498. Top panel:
radial prohle of observed C2H(l-0) intensity integrated over all
hyperhne components (jilled squares) and model prediction for a
constant abundance core with a central depletion hole (solid red
line). Middle panels: emerging spectra for each component of
the C^H^l-O) multiplet (black histograms) and predictions for
the same best fit model (solid red lines). Bottom panels: same
as middle panels but for the components of the C2H(2-1) mul-
tiplet. Note the reasonably good fit of all observables despite the
use of highly approximated collisional rates (see text).
3 -
L1498 Monte Carlo model
C 2 H(N=l-0)
10 1
2 10
R (cm)
3 10
Fig. 12: Radial prohle of excitation temperature for each hyper-
fine component of C2H(l-0) as predicted by our best fit Monte
Carlo model (each component is labelled according to the order-
ing in Table 3). Note the gradual drop of T ex with radius, which
is caused by the combined decrease in collisional excitation and
photon trapping towards the outer core. The different hyperhne
components have different T ex depending mostly on trapping ef-
fects (see text for a full discussion).
9 x 10 16 cm inside which the C2H abundance is negligible (10~ 4
times the outer value), similar to that of other species in L1498,
see Tafalla et al. 2006. The results of this model are shown in
Fig.QT](red lines) superposed to observations (black histograms
and squares). As it can be seen, the model fits reasonably well
both the radial prohle of N=l-0 intensity and the spectra of
most components in both the A^= 1 -0 and N=2— 1 multiplets. The
model in addition, fits reasonably well the observed ratios of line
pairs presented in Fig.|9]
If our Monte Carlo model fits the observed C2H emission,
we can use it to analyze the excitation of the different N= 1 -0
components and understand the origin of the observed line ra-
tios. This is best done by studying the radial prohle of excitation
temperature, which is presented in Fig.Q~2] As the hgure shows,
the r ex of each component systematically decreases with radius,
from about 6 K in the core interior to a value close to the cos-
mic background temperature near the core edge. The hgure, in
addition, shows that at each radius, the T ex of the different com-
ponents can differ by as much as 1 K, and that component 2
is signihcantly more excited than the rest. These differences in
excitation over the core and among components could in prin-
ciple result from differences in the contribution of collisions or
from photon trapping, and to disentangle the two effects we have
run an alternative model having a factor of 10 4 lower abundance.
This optically thin case also presents an outward drop in T ex , this
time entirely due to the effect of collisions, but it predicts exci-
tation temperatures in the core interior that are about 1 K lower
than for the best fit model. In addition, component 2 in this thin
case has a T ex comparable to that of the other components, and
the overall scatter of r ex among all components does not exceed
0.5 K. This decrease in the excitation when the lines become thin
indicates that in the best fit model the N = 1 level populations
are signihcantly enhanced by trapping of A^ =1-0 photons. The
higher excitation of component 2, in particular, appears as an
extreme case of trapping: this component has the largest relative
intensity and therefore is the most sensitive one to optical depth
effects. Its sensitivity to trapping makes it the brightest line of the
multiplet despite suffering from self absorption at the line cen-
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
11
tre. Differential line trapping seems also responsible for the dif-
ferent intensity of components 3 and 4, which would otherwise
be equally bright because of their equal Einstein A coerhcient
and upper level statistical weight. As Fig.[T2lshows, component
4 has an ~ 0.5 K higher T ex in the inner core, and this is most
likely the result of enhanced trapping due to an overpopulation
of the N,J,F = 0, 1/2, 1 level, where most N=l-0 transitions
end (except for components 3 and 6).
In summary, our model shows that most excitation "anoma-
lies" of the hyperhne components in the N =1-0 multiplet can
be explained as resulting from the different balance between col-
lisions and trapping expected for lines of very different intrinsic
intensities under conditions of subthermal excitation and high
optical depth. Further work on this issue requires an improved
set of collision rates for C2H, and we encourage collision rate
modelers to consider this species for future computations.
6. Column density and abundance estimate
Deriving a column density for C2H requires an estimate of the
excitation temperature. This as we see in the Monte Carlo calcu-
lations discussed above depends on the line transfer and hence
on position within the core. However a reasonable approxima-
tion to the column density can be obtained assuming a homoge-
neous layer with constant excitation temperature. We have veri-
fied this assumption for the case of L1498 using the Monte Carlo
program.
The excitation temperature r ex can be inferred (method 1)
from an LTE fit to the hyperhne components of either the N= 1 -0
or N=2— 1 lines assuming unity beam hlling factor and using the
measured intensity of optically thick transitions. An independent
measure can be obtained (method 2) from the ratio of intensities
of low line strength transitions of the N=2-l and N=l-Q lines
assuming them to be optically thin. We summarize in Table[7]the
excitation temperatures derived using these different approaches
for the positions where we have N=2-l data available. From
Table|7]we see that in both sources, the excitation temperature
appears to be between 3.8 and 4.9 K. We will assume an excita-
tion temperature of 4 K in the following for all positions.
the upper level, r ex the excitation temperature, j r m bdv the inte-
grated intensity, J(T) as dehned in eq.|4]and Q = }ZJ=o gje~ E i' kTa
is the partition function. The integral in equation |6] is over all
six hyperhne components. In practise, the stronger components
of the line are sometimes thick and we have hence integrated
only over the 87284 MHz and 87446 MHz lines (components 1
and 6 in Table|4]i and divided the result by the sum of their line
strengths (0.085) in order to evaluate the integral in equation|6]
We need to compare this with the molecular hydrogen col-
umn density A^(H 2 ) and we do this using the mm dust emission
(from Tafalla et al. 2004 for L1498 and from this work, see Sect.
12.31 for CB246). We assume here in both sources a dust grain
opacity of 0.005 cm 2 g -1 of dust and a dust temperature of 10 K,
as in Sect.|3~4l
We also evaluated the column densities and the abundances
of N 2 H + and CO for CB246 (using our data, see Sect. [2T2]> and
for L1498 (using data from Tafalla et al. 2004). We derived
A^(N 2 H + ) following the same procedure as for A^(C 2 H). Thus,
if the stronger components of N 2 H + (l-0) are thick then we in-
tegrated only over the isolated component (F[ ,F' — > Fi,F =
1 , 2 — > 0, 1 ), dividing the result by its line strength (0. 1 1 1 ). In the
case of C ls O, we used directly equation|6]to calculate A^(C 18 0)
and N(CO) assuming [ 16 O]/[ I8 O]~560 (Wilson & Rood 1994).
We give in Table [8] abundance estimates that we have made at
selected positions in L1498 and CB246.
Table 8: Molecular abundances in L1498 and CB246.
o-,5offsets [C 2 H]/[H,] [N 2 H + ]/[H 2 ] [CO]/[H 2 ]
[","]
[10
- 8 ]
[10- 10 ]
[10
- 5 ]
L1498
0,0
0.8 ±
0.1
1.9 ±0.4
0.5 ±
0.1
-40, 20
1.1 ±
0.3
1.8 ±0.5
0.8 ±
0.2
60, -40
1.5 ±
0.4
2.4 ± 0.7
1.1 ±
0.3
CB246
0,-15
1.0 ±
0.3
3.1 ± 1.6
0.9 ±
0.2
-60, 15
1.2 ±
0.5
2.4 ± 1.7
1.0 ±
0.3
-60, 60
1.0 ±
0.3
1.2 ±0.9
1.3 ±
0.4
-30, 15
1.0 ±
0.2
3.0 ± 1.2
1.6 ±
0.5
Table 7: Yalues of r ex [K] derived using the two methods.
a, S orTsets
method 1
method 2
[","]
C 2 H(l-0) C 2 H(2-1)
L1498
0,0
4.0 ±0.1 4.6 ±0.6
4.2 ±0.6
60,-40
4.1 ±0.1 4.3 ±0.8
4.4 ±0.5
CB246
0,0
4.1 ±0.1 4.9 ±1.0
4.4 ±0.9
15,-30
4.1 ±0.3 4.4 ±0.7
4.4 ± 0.4
-60,15
4.0 ±0.2 3.9 ±0.4
3.8 ±0.6
For an optically thin C 2 H(l-0) line, one can derive the col-
umn density, A^(C 2 H), through the formula:
N(C 2 H)
&nA Q
„Ej/kT a
j T mb dv
C 3 Ajigj e^-l^ - 1 J(T ex ) - J(T hg )
(6)
where v ; , is the transition frequency, c the speed of light, A « the
Einstein coefficient, gj the statistical weight, Ej the energy of
In Fig. [13] we show the plot of C 2 H, N 2 H + and CO col-
umn density against H 2 column density at different positions in
both sources. We find that for column densities A^(H 2 )< 2 x 10 22
cirT 2 , N(C 2 H) is proportional to N(U 2 ) but, in L1498, for higher
values of A^(H 2 ), this proportionality breaks down and N(C 2 H)
seems to saturate at a value of ~ 2.2 x 10 14 cirT 2 . This is likely
due to depletion of C 2 H in the high density core of L1498 as al-
ready suggested by the maps in Fig.[JJ the cuts in Fig.[3]and the
Monte Carlo modeling. This behaviour is in complete analogy
with several other species including the carbon rich molecule
C 3 H 2 (Tafalla et al. 2006).
On the other hand, for A^(H 2 ) less than 2 x 10 22 cirr 2 , we
find that the ratio N(C 2 H)/N(H 2 ) is constant corresponding to a
constant C 2 H abundance in the low density part of the core. This
corresponds to an average estimated C 2 H abundance relative to
H 2 of (1.0+0.3) x 10~ 8 inL1498 and (0.9 +0.3) x 10~ 8 in CB246.
Our C 2 H abundances estimate for L1498 using eq. |6]agrees to
within 20% with the Monte-Carlo estimate discussed in Section
15.21 It is noteworthy and somewhat surprising to us that the C 2 H
abundances are so similar in cores of differing characteristics.
12
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
A more strict correlation over the whole range of observed
H2 column density is shown by the ratio Af(N 2 H + )/Af(H 2 ) which
corresponds to an average estimated N 2 H + abundance relative
to H 2 of (1.8 + 1.2) x lfr 10 in L1498 and (1.8 ± 0.8) x 10" 10
in CB246. C ls O behaves differently showing a rather constant
value of its column density over the A^(H 2 ) range corresponding
to Af(CO)=(2.0+0.4)x 10 17 cirT 2 in CB246 and (2.1 ±0.3)x 10 17
cm~ 2 in L1498, while the average abundance relative to H2 is
(1.3 ± 0.6) x 10~ 5 in L1498 and (1.5 ± 0.4) x 10~ 5 in CB246.
One dirTerence between CB246 and L1498 is that the N(U 2 )
peak (see Fig. [TSl is a factor 1.33 smaller in CB246 and in-
deed does not reach values for which large depletion is noted
in L1498. However, our map in C ls O(2-l) (see lower panel of
Fig.|2]) as well as the cuts shown in Fig.|6]suggest to us that some
CO depletion does occur in CB246.
7. Discussion
In this study, we have compared two cores of rather different
characteristics. L1498 is embedded in the Taurus complex with
a total dimension of 2.6 x 2.0 arc minutes (Caselli et al. 2002)
corresponding to 0. 10 x 0.08 parsec at the distance of 140 parsec
(Elias, 1978). CB246 is an isolated globule with dimensions of
7.8 X 4.5 arc minutes corresponding to 0.45 x 0.26 parsec at our
adopted distance of 200 parsec. It is thus perhaps not surprising
that the non-thermal component of the line width, obtained from
the observed line widths corrected for thermal broadening (see
e.g. Caselli & Myers 1995), is 0.23 km s" 1 in CB246 as com-
pared to 0.14 km s _1 in L1498. However, apart from that, many
characteristics of these two objects seem similar as discussed
below.
7.1. Non-LTE effects
One of our aims was to test for non-LTE effects in C2H if any in
the hope that this could lead to a better understanding of physi-
cal conditions in cores. Our results show that non-LTE effects do
occur and they are qualatively similar in the two cores. With our
Monte Carlo radiative transfer program, we have shown that one
can qualitatively understand their nature in terms of trapping of
individual components of the AT=1— and A^=2-l lines. These
calculations also show that reasonable estimates of parameters
such as the C2H column density can be obtained using a simple
one-layer homogeneous model with constant excitation temper-
ature. Given that real cores are not spherically symmetric, this is
a useful simplihcation. Clearly, more rehned models will even-
tually be useful but equally clearly, this requires computations of
collisional rates for C2H along the lines of those carried out by
Monteiro and Stutzki (1986) for HCN.
7.2. Chemistry
There has been signihcant recent work on the chemistry of
prestellar cores (see e.g. Aikawa et al. 2003, 2005, Lee et
al. 2004, Flower et al. 2006, Akyilmaz et al. 2007). These stud-
ies follow the evolution of molecular abundances in a collapsing
prestellar core as a function of time and initial conditions. For
example, the Aikawa et al. (2005) model, a model with a, the
ratio of gravitational to pressure force, equal to 1.1, predicts a
[C 2 H]/[H 2 ] abundance ratio of 4 x 10~ 9 at a radius of 10000
AU (i.e. outside the depletion hole) or a factor of roughly 3
smaller than our observed value but there are a variety of fac-
tors which induence this estimate as we now discuss. One factor
5 10 15 20 25 30 35
N(H 2 ) [10 21 cm" 2 ]
Fig. 13: Plot of the column density of C 2 H, N 2 H + and CO against
the H 2 column density for L1498 (open squares) and CB246
(jilled sguares). The dotted line represents the 2<x = 6 mJy limit
of the continuum emission. Data for N 2 H + and CO for L1498
are from Tafalla et al. (2004).
is that results are sensitive to the degree of depletion onto grain
surfaces which, in turn, depends on the effective grain cross sec-
tion. Another important parameter is the effective [C]/[0] ratio in
the gas phase because this is critical for the formation of C-rich
species such as C3H 2 , HC3N, and C 2 H (see Terzieva & Herbst
1998, Akyilmaz et al. 2007). A [C]/[0] ratio close to unity can
occur when a large fraction of the available oxygen condenses
out in the form of water ice on grain surfaces. In this situation,
the model calculations suggest that CO takes the lion's share of
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
13
both carbon and oxygen in the gas phase and that, as a result,
little oxygen is left in atomic or molecular form.
The results of Terzieva & Herbst (1998) show in particular
that the C2H abundance can change by an order of magnitude
due either to the assumptions concerning [C]/[0] or due to as-
sumptions concerning the rates of certain neutral-neutral reac-
tions. We find in contrast that the observed C2H abundance does
not seem to change greatly as a function of position and in fact,
to a reasonable approximation, C2H follows the C 18 distribu-
tion. Where CO appears to be depleted in L1498, C2H seems
to be depleted also and this also seems to be true for CB246 al-
though the depletion holes are smaller. To some extent, we pre-
sume that this reAects the fact that CO is the main source of gas
phase carbon in cores. When CO depletes, there is less carbon
available to form species with 2 or more C atoms. It is strik-
ing also that we derive very similar C2H abundances in the two
cores which we have studied. This may be due to chance but it is
interesting that our estimated C2H abundance is similar to that
derived in TMCl and L183 (see Table 2 of Terzieva and Herbst
1998).
It is also worth noting that C2H, being a radical, is subject
to destruction by atomic oxygen and nitrogen. Thus from the
UMIST database (Woodall et al. 2007), one nnds that C 2 H re-
acts with O forming CO and CH at a rate of 1 .7 x 10" n cm 3 s _1
and the rate for atomic N is similar. This compares with a rate for
depletion onto dust grain surfaces of roughly 10~ 17 nH s _1 where
ne is the hydrogen nucleon density in crrr 3 and a rate of 10~ 9
cm 3 s _1 for reactions with ions such as C + and H + . For an ion-
ization degree of 10~ 8 , which is fairly typical in cores (see Fig. 5
in Walmsley et al. 2004), reactions with ions and depletion onto
grain surfaces are competitive with one another. Destruction by
atomic oxygen and nitrogen will dominate if the abundances of
these species exceeds about 6 x 10~ 7 relative to hydrogen. The
atomic abundances are hence critical for the C 2 H abundance and
it is presumably for this reason (at least as far as O is concerned)
that the results of model calculations show C2H to be very sensi-
tive to the gas phase [C]/[0] ratio (e.g. Terzieva and Herbst 1998,
their Table 5). In fact, the only one of their model predictions for
C2H which compares reasonably with our observed values has
a [C]/[0] ratio of 0.8 (run 3 in their Table 5). We conclude that
high [C]/[0] gas phase ratios (and consequently a low atomic
O abundance) are part of the explanation of the relatively high
observed C2H abundance. This presumably implies a scenario
with a large fraction of oxygen locked-up in the form of water
ice on grain surfaces (see e.g. Hollenbach et al. 2009).
for L1498 as well as the expected Zeeman splitting factors Z, in
Hz fiG~ l , from Bel and Leroy (1998) to derive expected values
of cr B for all six C2H(l-0) transitions. The results, for an hour
of integration, are given in Table [9] where we see component 5
of C2H is the most favorable for detecting Zeeman splitting and
that, as for CN, there is a large variation of expected splitting
between components. This latter property is of fundamental im-
portance for the purpose of eliminating instrumental effects such
as beam squint.
Table 9: Expected values of c B for all six C2H(l-0) transitions
in L1498 at the offset (60, -40) for an hour of integration and a
spectral resolution equal to 40 kHz.
comp.
Z
r .
Avfwhm
0"B
no.
[Hz yuG- 1 ]
[K]
[kHz]
IpG]
1
2.6
0.9
80
740
2
0.7
1.4
100
2210
3
2.3
1.0
90
860
4
0.93
1.1
95
1990
5
2.8
1.1
85
590
6
0.93
0.9
75
1860
See Table|4]for component labels.
One sees also that without a considerable improvement in
system temperature, it will be diSicult to get below 3cr B limits of
200 nG. There have been claims of helds of this order however
(e.g. Shinnaga et al. 1999); Crutcher (1999) found a reasonable
fit to the Zeeman data available at the time of Z?i os = 80 x n^ 46
fiG with «4 = 10~ 4 «(H2) which would suggest that helds of
order 200 fiG are reasonable in sources like L1498. We conclude
therefore that helds of this order may be detectable using C2H
but it is extremely difficult with current sensitivities. On the other
hand, in more general, prospects do not seem worse than with
CN.
8. Conclusions
We have carried out a study of the behaviour of the abundance
of C2H toward two starless cores with contrasting properties,
L1498 and CB246. The main conclusions of our work are the
following:
7.3. C2H as a magnetic field probe
As mentioned in the Introduction, C2H is potentially capable of
being used to measure the Zeeman effect. The CN(l-O) transi-
tion has been successfully used for this purpose (Falgarone et
al. 2008) and superhcially at least, C2H has similar characteris-
tics to CN (they are iso-electronic). We have therefore used the
results of Crutcher et al. (1996) together with the Zeeman split-
ting calculations of Bel & Leroy (1998) to infer the expected
RMS sensitivity to magnetic field crg of a C2H measurement at
the intensity peak in L1498 for an integration time t int . Thus we
take
2 T sys AypwHM
o"B~-^r^— — (7)
Z T A 6vt inl
In this equation, we use our observed line intensities T* A , system
temperature r sys , the spectral resolution 5v and width Avfwhm
1. In L1498, C2H shows a distribution similar to that observed
in other species attributed to depletion onto grain surfaces
in the central region; by contrast, in CB246 the dust and the
C2H emission have similar distributions.
2. The two cores show a clear signature for deviations from
LTE populations in C2H: spectra in most positions devi-
ate from expectations assuming LTE and thus a single-
temperature LTE model cannot fit the data.
3. There are positions of high C2H optical depth in both
sources; in addition, L1498 shows self-absorption toward the
dust peak.
4. Our Monte Carlo model shows that the observed deviations
from LTE can be qualitatively understood, but reliable colli-
sional rate calculations for C2H are needed in order to make
furtherprogress.
5. The non-LTE deviations have not prevented from computing
column density values based on LTE. We also found that the
C2H abundance relative to H2 is remarkably constant outside
14
M. Padovani, C.M. Walmsley, M. Tafalla, D. Galli and H.S.P. Miiller: C 2 H in prestellar cores
regions of high CO depletion with a value of (1 .0+0.3)xl0
in L1498 and (0.9±0.3)xl0 8 in CB246. One possible impli-
cation is that the abundances of atomic oxygen and nitrogen
are extremely low (below 6 x 10~ 7 relative to H).
6. We derived a new set of frequencies for all the six hy-
pernne components of C2H(l-0) and seven components of
C2H(2-1), computing an improved set of spectroscopic con-
stants for C2H.
Acknowledgements. This effort/activity is supported by the European
Community Framework Programme 7, Advanced Radio Astronomy in Europe,
grant agreement no.: 227290. M.P. and D.G. acknowledge support from the
EC Research Training Network MRTN-CT-2006-035890 "Constellation:
The Origin of Stellar Masses". H.S.P.M. is grateful tbr support by the
Bundesministerium fiir Bildung und Forschung (BMBF) administered through
Deutsches Zentrum tur Luft- und Raumtahrt (DLR). His support is aimed in
particular at maintaining the CDMS.
Shirley, Y. L., Nordhaus, M. K., Grcevich, J. M., Evans, N. J., II, Rawlings, J.
M. C. & Tatematsu, K. 2005, ApJ, 632, 982
Tafalla, M., Myers, P. C, Caselli, P., Walmsley, C. M. & Comito, C. 2002, ApJ,
569, 815
Tafalla, M., Myers, P. C, Caselli, P. & Walmsley, C. M. 2004, A&A, 416, 191
Tafalla, M., Santiago-Garcia, J., Myers, P. C, Caselli, P, Walmsley, C. M. &
Crapsi, A. 2006, A&A, 455, 577
Tassoul, J.-L. 1978, Theory of Rotating Stars (Princeton: Princeton Univ. Press)
Terzieva, R. & Herbst, E. 1998, ApJ, 501, 207
Thum, C, Wiesemeyer, H., Paubert, G., Navarro, S. & Morris, D. 2008, PASP,
120, 777
Turner, B. E., Terzieva, R. & Herbst, E. 1999, ApJ, 518, 699
Walmsley, C. M., Flower, D. R. & Pineau des Forĕts, G. 2004, A&A, 418, 1035
Woodall, J., Agundez, M., Markwick-Kemper, A. J. & Millar, T. J. 2007, A&A,
46,1197
Reterences
Aikawa, Y., Ohashi, N. & Herbst, E. 2003, ApJ, 593, 906
Aikawa, Y, Herbst, E., Roberts, H. & Caselli, P. 2005, ApJ, 620, 330
Akyilmaz, M., Flower, D. R., Hily-Blant, P, Pineau Des Forĕts, G. & Walmsley,
C.M. 2007, A&A, 462, 221
Andrĕ, P, Ward-Thompson, D. & Motte, F. 1996, A&A, 314, 625
Bel, N. & Leroy, B. 1998 A&A, 335, 1025
Bergin, E. A. & Tatalla, M. 2007, Ann. Rev. A&A, 45, 339
Caselli, P, Myers, P. C. & Thaddeus, P. 1995, ApJ, 455, L77
Caselli, P, Benson, P. J., Myers, P. C. & Tafalla, M. 2002, ApJ, 572, 238
Chandrasekhar, S. & Miinch, G. 1950, ApJ, 111, 142
Clemens, D. P. & Barvainis, R. 1988, ApJS, 68, 257
Codella, C. & Scappini, F. 1998, MNRAS, 298, 1092
Crutcher, R. M., Troland, T. H., Lazareff, B. & Kazes, I. 1996, ApJ, 456, 217
Crutcher, R. M. 1999, ApJ, 520, 706
Dame, T. M., Ungerechts, H., Cohen, R. S., de Geus, E. J., Grenier, I. A., May,
J., Murphy, D. C, Nyman, L.-A. & Thaddeus, P. 1987, ApJ, 322, 706
Ebenstein, W. L., & Muenter, J. S. 1984, J. Chem. Phys., 80, 3989
Elias, J. H. 1978, ApJ, 224, 857
Flower, D. R., Pineau Des Forĕts, G. & Walmsley, C. M. 2006, A&A, 456, 215
Falgarone, E., Troland, T. H., Crutcher, R. M. & Paubert, G. 2008, A&A, 487,
247
Goodman, A. A., Benson, P. J., Fuller, G. A. & Myers, P. C. 1993, ApJ, 406, 528
Gottlieb, C. A., Gottlieb, E. W. & Thaddeus, P. 1983, ApJ, 264, 740
Green, S. & Thaddeus, P. 1974, ApJ, 191, 653
Guilloteau, S. & Baudry, A. 1981, A&A, 97, 213
Hily-Blant, P, Walmsley, C. M., Pineau des Forĕts, G. & Flower, D. 2008, A&A,
480, L5
Hollenbach, D., Kautman, M. J., Bergin, E. A. & Melnick, G. J. 2009, ApJ, 690,
1497
Kirk, J. M., Ward-Thompson, D. & Crutcher, R. M. 2006, MNRAS, 369, 1445
Launhardt, R. & Henning, T. 1997, A&A, 326, 329
Lee, J.-E., Bergin, E. A. & Evans, N.J. II 2004, ApJ, 617, 360
Lemme, C, Wilson, T. L., Tiertrunk, A. R. & Henkel, C. 1996, A&A, 312, 585
Lique, E, van der Tak, F. F. S., Klos, J., Bulthuis, J. & Alexander, M. H. 2009,
A&A, 493, 557
Monteiro, T. S. & Stutzki, J. 1986, MNRAS, 221, P33
Miiller, H. S. P, Klaus, T. & Winnewisser, G. 2000, A&A, 357, L65
Miiller, H. S. P, Thorwirth, S., Roth, D. A., & Winnewisser, G. 2001, A&A, 370,
L49-L52
Miiller, H. S. P, Schloder, F, Stutzki, J., & Winnewisser, G. 2005, J. Mol. Struct,
742, 215
Myers, P. C, Mardones, D., Tafalla, M., Williams, J. P. & Wilner, D. J. 1996,
ApJ, 465, L133
Myers, P. C. & Benson, P. J. 1983, ApJ, 266, 309
Ohashi, N. 1999, in Proceedings of Star Formation 1999, ed. T. Nakamoto
(Nobeyama Radio Observatory), 129
Pagani, L., Bacmann, A., Cabrit, S. & Vastel, C. 2007, A&A, 467, 179
Pagani, L., Daniel, F. & Dubernet, M.-L. 2009, A&A, 494, 719
Preibisch, Th., Ossenkopf, V, Yorke, H.W. & Henning, Th. 1993, A&A, 279,
577
Sastry, K. V. L. N., Helminger, P, Charo, A., Herbst, E., & De Lucia, F. C. 1981,
ApJ, 251,L119
Schmid-Burgk, J., Muders, D., Miiller, H. S. P, & Brupbacher-Gatehouse, B.
2004, A&A, 419, 949
Shinnaga, H., Tsuboi, M. & Kasuga, T. 1999, in Proceedings of Star Formation
1999, ed. T. Nakamoto (Nobeyama Radio Observatory), 175